Those of ordinary skill in the field of semiconductor devices will be familiar with many different types of such devices, including, for example, microprocessors and various types of memory devices, such as dynamic random-access memory devices ("DRAMs"), synchronous DRAMs ("SDRAMS"), static random-access memory devices ("SRAMS") capable of storing millions of bits of digital information.
Those of ordinary skill in the art will be familiar with the practice of designing and implementing semiconductor devices which are capable of being permanently programmed during the fabrication process to exhibit different operational properties depending upon a selection process performed at some point during manufacture or testing of the device. Such a practice is common in connection with the design and manufacture of semiconductor memory devices. For example, it is common for a semiconductor memory device to be designed such that during or after the fabrication process, the manufacturer has the option of selecting one of a plurality of input/output (I/O) configurations for the device. A memory device having a (nominally) 64 megabit capacity may be configured to have one of several I/O configurations: either a 16 megabit-by-four-bit (".times.4") I/O configuration, where each row and column address pair references four bits at a time, or a 8 megabit-by-eight-bit (".times.8") I/O configuration, where each row and column address pair references eight bits at a time, or a 4 megabit-by-sixteen-bit (".times.16") configuration, where each row and column address pair references sixteen bits at a time. The selection of either the .times.4, .times.8, or the .times.16 I/O option commonly involves the actuation ("blowing") of one or more one-time-programmable devices on the semiconductor device. Once the appropriate programmable devices are actuated, the device thereafter will permanently operate in accordance with the selected I/O configuration.
So-called "antifuses" are often used as the one-time programmable devices in an integrated circuit product as a mechanism for changing the operating mode of the product. That is, anti-fuses are often used for the purpose of permitting the selection from among a plurality of programmable options for a semiconductor device. As those of ordinary skill in the art will appreciate, antifuses are essentially one-time programmable switching devices whose conductivity state (conductive or non-conductive) can be altered through application of predetermined programming signals to an integrated circuit's I/O pins. Most commonly, an antifuse is initially (i.e., at the time of fabrication) "open" or non-conductive. If it is desired to actuate or "blow" a particular antifuse to change an operational characteristic of the integrated circuit, predetermined programming signals are applied to the integrated circuit's I/O pins. Once blown, the antifuse is rendered conductive. Further, once blown, it is typically not possible to reverse the programming. That is, once a fuse has been blown, it cannot be un-blown.
Programmable options such as the I/O configuration of a memory device are often referred to as "fuse options" for the device. Those of ordinary skill in the art will appreciate that the I/O configuration of a memory device is but one example of the type of fuse options that may be available for a particular device. Fuse options may be available in connection with many different operational parameters of a semiconductor device, including without limitation the selection of certain internal timing parameters, the availability and activation of redundant rows or columns of memory cells in a memory device, the operational speed of a device, voltage regulation of a device, and so on. Providing a single semiconductor device with one or more fuse options is regarded as desirable, since a single design and fabrication process can be used to manufacture more than one class of end product. This flexibility eliminates the need for separate designs and separate fabrication processes to produce multiple classes of end product. Additionally, fuse options enable the manufacturer to counteract the effects of semiconductor process variations, advantageously increasing fabrication yield and maximizing production of higher-performance parts.
A perceived limitation on the fuse option programmability of an integrated circuit arises from the one-time nature of antifuse programming. That is, once a particular fuse option has been selected to change the operational mode of a given integrated circuit, it is typically not possible to reprogram the integrated circuit back to its original operational mode. This limitation may be undesirable. For example, in some cases it may not be possible to ascertain whether a given operational mode is appropriate for an integrated circuit until the integrated circuit can be observed operating in that mode. However, due to the one-time programmability of antifuses, operational parameters cannot be tentatively selected and subsequently unselected.
Semiconductor devices or circuits for providing reprogrammable option selection in an integrated circuit are known. For example, a floating-gate transistor which can be repeatedly and alternately rendered conductive or non-conductive (i.e., turned on and off) can be used to provide option reprogrammability for an integrated circuit. However, such devices have certain potential disadvantages. Circuits or devices for providing option programmability and reprogrammability can be more difficult to fabricate than antifuses, occupy more area in the integrated circuit, increase the overall power consumption of the integrated circuit, and be more difficult to program than antifuses.